NeuCo has been working with the U.S. Department of Energy (“DOE”) under the Clean Coal Power Initiative (CCPI) program for almost a decade.  Our first project, which took place at Dynegy’s Baldwin Energy Complex in Baldwin, IL, was completed in 2007.  In 2006, NeuCo also began working on a Round 2 Mercury Specie and Multi-Pollutant Control Project at NRG’s Limestone Generating Station in Jewett, Texas.  

After years of intensive efforts from NeuCo, Limestone, DOE and vendor partners, the Round 2 project recently concluded with the publishing of the final technical report.  The project’s intent was to demonstrate the ability to affect and optimize mercury speciation and multi-pollutant control using non-intrusive advanced sensor and optimization technologies.  Final analysis concluded that the plant significantly reduced NOx emissions, improved heat rate, successfully changed its fuel blend to reduce Mercury emissions, and instituted a streamlined equipment anomaly detection and diagnostics system.

Rob James, Product Manager, was the NeuCo project lead on both CCPI projects. We got a chance to sit down with him and ask a few questions.  

How does the CCPI Round 2 project at Limestone compare to the CCPI Round 1 project at Baldwin?
The challenges presented by Round 1 at Baldwin and Round 2 at Limestone were completely different. At Baldwin we were essentially developing version 2 of the ProcessLink® platform and building three new product applications on it at the same time; Limestone was about unit-wide optimization and incorporated a lot of third-party instrumentation. Limestone was a huge integration challenge that required more people and systems to work together.

Any lessons learned from Round 1 to Round 2?
From Round 1 to Round 2, ProcessLink continued to improve and evolve such that by Round 2 we were able to attack a much more diverse set of problems, use more advanced approaches and be more efficient about it.

What are some of the things about the Limestone project you are most proud of?
Given the ambitiousness of this project and the huge number of moving parts, I’m proud of how we were able to adapt to the changing set of constraints in a way that provided a lot of utility and met a key set of investigative goals.  For instance, we had to respond to changing economic drivers, such as the changing cost-benefit relationship between NOx and heat rate and changes in the value of NOx and SO2 credits, as well as a range of equipment and instrumentation constraints. Reliably measuring Mercury in multiple locations, over a long period of time, was also a major challenge.

We felt it was really important to balance the needs and wishes of all project investors. While we were asking a number of important investigative questions and searching through a large array of uncertainties, we really wanted to be useful to the plant in the short-term and in a concrete way. I believe we were able to do that. That is what our stuff is good at – attacking the combined challenge of discovery and value delivery.

One of the specific outputs of the project I’m most proud of is ProcessLink’s enhanced Optimization Measurement Tool, which structures an interactive on-line benchmarking process. I believe this is one of the main requirements for operationalized AI. Because benchmarking is so challenging in complex, noisy situations (as are often found in this industry), it’s often either dismissed as infeasible or approached naively. With optimization we do not have that luxury; we really need to be able to define and quantify  (i.e. measure) simultaneous change in multiple related dimensions in order to apply some real notion of “good” and “bad” to how things are going.  Without a tool to help do this, there’s no way I could have developed a data-driven report that addressed the significant questions the project sought to answer.

What were some of the biggest challenges?
Integrating all of the different types of instrumentation was harder than we thought. For instance it was really difficult to get reliable Mercury measurements for use by the optimizer in the places we wanted them – i.e. at the ESP inlet, FGD inlet and FGD outlet. These are just tough places to put sensitive instruments and analyzer packages.

You worked closely on this project with Steve Piche, NeuCo’s Director of Research & Development. Could you discuss his contribution?
Steve was involved in all aspects of the project. He is a real live math freak and one of my favorite teachers on this and many other projects. Steve is especially knowledgeable about modeling and process chemistry, and was solely responsible for the development of the Hg VOAs (Virtual On-line Analyzers), without which we could not have done the project. It means a lot to me to work with guys like Steve.

Of course, many others at NeuCo contributed a tremendous amount to the project as well – not to mention all the people from the plant that we were able to learn a lot from.   

Any fond memories of your trips to Jewett, TX that you’d care to share with us?
One blazing hot afternoon I was driving to the plant and a big thunderstorm popped up.  I pulled off the main road to watch it hit. I had the world to myself for awhile and was expecting a tornado to appear any minute. The Texas countryside is really beautiful when the storms roll in.

Jewett also has at least one great hole-in-the wall Mexican restaurant that makes excellent fajitas. A number of guys from the Austin office can vouch for those fajitas.

Another cool thing about Jewett – and I only recently found this out – it’s the birthplace of Marine Corps Sergeant Romus (RV) Burgin, whose experiences on Peleliu and Okinawa during WWII were dramatized in the HBO Miniseries The Pacific, which I really appreciated.

How many rounds of edits did you and the DOE trade before agreeing on the final report?
I think it was five. That’s less than the nine rounds for the first CCPI project – probably because the Limestone report was more than 300 pages long…

www.neuco.net/blog